Dynamic monitoring system for geological stability of mining area

Abstract

The invention relates to the technical field of mining area geological monitoring, and discloses a dynamic monitoring system for mining area geological stability, which comprises a data acquisition module used for acquiring geological monitoring data of a plurality of data sources in a mining area; the data preprocessing module is used for performing time synchronization, space alignment, data cleaning and standardization processing on the acquired data; the long-period trend analysis module is used for carrying out time sequence modeling and multi-scale change rate analysis on the preprocessed data; and the adaptive risk model module is used for setting and dynamically adjusting a risk threshold based on the long-period trend analysis result. By introducing a long-period trend analysis and change rate evaluation technology, the recognition precision and early warning foresight of the slow change of the geological structure in the mining area are remarkably improved, and long-term accumulated change characteristics such as settlement and inclination can be captured; meanwhile, the self-adaptive adjustment of the alarm threshold is realized by using the dynamic risk model, so that the system can sensitively deal with geological states at different stages.

Description

Technical Field

[0001] The invention relates to the technical field of mining area geological monitoring, in particular to a dynamic monitoring system for geological stability in mining areas. Background Art

[0002] With the growing global demand for mineral resources, mining activities have expanded rapidly in depth and breadth. However, frequent mining and engineering activities in mining areas will cause various changes in geological structures, including surface subsidence, slope sliding, crack expansion, rock mass displacement, etc. These geological changes not only affect the safety of operations in the mining area, but also cause potential harm to the surrounding environment and ecosystem, especially in some large mining areas, where the geological stability problem is becoming increasingly serious.

[0003] Current geological monitoring systems often rely on a single data source, such as GNSS, InSAR and other equipment, through which displacement or settlement data is obtained. However, the monitoring range of a single data source is limited, and it is difficult to fully reflect the complex geological structure changes in the mining area. In addition, this type of monitoring system lacks the ability to analyze long-term geological changes, and can usually only perform static analysis based on short-term data, and cannot identify those slowly accumulating trends. For example, the slow settlement or tilt of the slope or underground structure may not be obvious in the short term, but it will significantly increase the risk over time. The short-term analysis of the existing system is difficult to capture this gradually accumulated risk signal in time, and thus cannot provide early warning.

[0004] In terms of risk threshold setting, traditional geological monitoring systems generally adopt a fixed threshold mode, setting alarm thresholds based on historical experience or benchmark values. However, the geological environment in mining areas is complex and changeable, and the geological state may change dynamically with changes in mining progress, stratigraphic conditions, climate and other factors. Fixed threshold modes are difficult to adapt to such complex environments, often leading to false alarms and missed alarms. Specifically, when geological conditions change, fixed thresholds may trigger alarms before the actual risk increases, resulting in unnecessary intervention; or when the risk rises sharply, the threshold cannot be adjusted in time, missing the best time to alarm. In addition, the setting of fixed thresholds requires frequent manual intervention, which reduces the degree of automation of the monitoring system, increases the burden of management, and cannot meet the needs of modern mining areas for intelligent and real-time early warning. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention provides a dynamic monitoring system for the geological stability of mining areas, which solves the problems that the geological monitoring system in mining areas cannot effectively identify slowly accumulated long-period geological changes, it is difficult to adjust the alarm threshold in time to cope with dynamic geological environment changes, and false alarms and missed alarms are frequent.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a dynamic monitoring system for geological stability in mining areas, comprising:

[0007] A data acquisition module for collecting geological monitoring data from multiple data sources within the mining area;

[0008] Data preprocessing module, used for time synchronization, spatial alignment, data cleaning and standardization of collected data;

[0009] Long-term trend analysis module, used to perform time series modeling and multi-scale rate of change analysis on preprocessed data;

[0010] Adaptive risk model module, used to set and dynamically adjust risk thresholds based on long-term trend analysis results;

[0011] A space-time data fusion module to fuse data from different sources and generate a geological risk index;

[0012] Abnormal event feedback and adaptive learning module, used to adjust risk threshold strategy based on monitoring result feedback;

[0013] The intelligent early warning and emergency response module is used to generate early warning signals and trigger emergency responses based on geological risk indexes.

[0014] Preferably, the data acquisition module includes:

[0015] GNSS monitoring equipment, used to collect three-dimensional displacement data;

[0016] InSAR data acquisition equipment, used to collect surface subsidence data;

[0017] Inclinometers, used to collect tilt angle data in mining areas;

[0018] Seismic sensors are used to collect data on microseismic activity.

[0019] Preferably, the data preprocessing module includes:

[0020] A time synchronization unit, used to align different types of data to a unified time step;

[0021] The spatial alignment unit is used to unify the data of different monitoring points into the same coordinate system;

[0022] The data cleaning unit is used to clean the data and remove outliers using the quantile regression method;

[0023] Data normalization unit, used to normalize data to the range [-1,1].

[0024] Preferably, the long-term trend analysis module includes:

[0025] The space-time joint decomposition unit is used to decompose the data into time trend terms, spatial variation terms and residual terms;

[0026] Wavelet transform multi-scale analysis unit is used to capture geological changes at different time scales.

[0027] Preferably, the adaptive risk model module includes:

[0028] An adaptive threshold setting unit, used to set an initial threshold based on the long-term trend analysis result;

[0029] The event-driven dynamic threshold update unit is used to dynamically adjust the risk threshold based on event feedback information when a geological anomaly event is detected.

[0030] Preferably, the dynamic adjustment in the event-driven dynamic threshold update unit is based on the following update formula:

[0031] T new =T prev +λ·ΔX

[0032] Among them, T new is the updated threshold, T prev is the previous threshold, λ is the adjustment coefficient, and ΔX is the change amplitude.

[0033] Preferably, the space-time data fusion module uses a Gaussian process regression method to fuse different data sources to generate a geological risk index I for the mining area. risk , the Gaussian process regression method is based on the following formula:

[0034] f(x)~GP(m(x),k(x,x′))

[0035] Among them, m(x) is the mean function, which is used to represent the average value of the overall geological stability, and k(x,x ′ ) is the covariance function, which is used to reflect the spatial-temporal similarity.

[0036] Preferably, the abnormal event feedback and adaptive learning module includes:

[0037] An abnormal event detection unit, used for detecting abnormal events based on local abnormal factors;

[0038] The adaptive feedback learning unit is used to use the Ql earning reinforcement learning method to adjust the risk threshold strategy according to the feedback of the detected abnormal events.

[0039] Preferably, the wavelet transform multi-scale analysis unit performs wavelet decomposition based on the following formula:

[0040]

[0041] Among them, W ψ (a,b) are wavelet coefficients, a is the scale factor, and b is the translation factor.

[0042] Preferably, the intelligent early warning and emergency response module includes:

[0043] a risk assessment unit for receiving the geological risk index and comparing it with a dynamic threshold to determine the geological risk level;

[0044] A multi-level warning generation unit, used to generate corresponding warning signals according to the geological risk level;

[0045] An emergency decision-making unit, used to generate an emergency response plan corresponding to the warning signal;

[0046] A dynamic emergency response optimization unit, used to adjust the emergency response plan according to actual geological change information;

[0047] Emergency plan simulation and drill unit, used to simulate and drill emergency plans;

[0048] The log recording and report generation unit is used to record information during the early warning and emergency response process and generate event reports.

[0049] The present invention provides a dynamic monitoring system for geological stability in mining areas. It has the following beneficial effects:

[0050] 1. The present invention significantly enhances the ability to identify changes in geological structures in mining areas by introducing long-term trend analysis and change rate assessment technology. By using space-time joint decomposition and wavelet transform and other methods, the system can capture the slowly accumulated change characteristics of long-term settlement, tilt and other geological data. This trend analysis not only improves the accuracy of identifying slow deformation risks, but also has foresight. It can issue early warning signals in the process of gradual accumulation of geological risks, reminding managers to take measures in advance, thereby effectively preventing the occurrence of sudden disasters. At the same time, the method is highly adaptable and can be used in different monitoring scenarios, such as slope slip monitoring and underground structure settlement assessment, providing a reliable trend analysis tool for different geological environments in mining areas.

[0051] 2. The present invention realizes adaptive adjustment of alarm thresholds through dynamic risk models. Traditional fixed threshold modes often cannot effectively adapt to real-time changes in geological conditions, while dynamic risk models can automatically adjust alarm thresholds based on historical monitoring data and real-time data to ensure that the system can sensitively respond to geological conditions at different stages. This adaptive adjustment mechanism reduces false alarms and missed alarms while also improving the intelligence of the early warning system. The system can automatically optimize threshold settings based on geological stability, increase thresholds when risks are low to avoid frequent alarms, and lower thresholds when risks increase to trigger alarms in a timely manner, thereby achieving higher response efficiency when responding to sudden risks. In addition, the automated threshold adjustment mechanism reduces the need for manual intervention, making the system run more efficient and reliable, and providing an intelligent geological risk management solution for mining areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a framework diagram of the system of the present invention;

[0053] Figure 2 It is a framework diagram of the data acquisition module of the present invention;

[0054] Figure 3 It is a framework diagram of the data preprocessing module of the present invention;

[0055] Figure 4 It is a framework diagram of the long-term trend analysis module of the present invention;

[0056] Figure 5 It is a framework diagram of the adaptive risk model module of the present invention;

[0057] Figure 6 It is a framework diagram of the abnormal event feedback and adaptive learning module of the present invention;

[0058] Figure 7 It is a framework diagram of the intelligent early warning and emergency response module of the present invention. DETAILED DESCRIPTION

[0059] The following will be combined with the drawings in the specification of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0060] Please see attached Figure 1 , an embodiment of the present invention provides a dynamic monitoring system for geological stability of a mining area, comprising:

[0061] A data acquisition module for collecting geological monitoring data from multiple data sources within the mining area;

[0062] Data preprocessing module, used for time synchronization, spatial alignment, data cleaning and standardization of collected data;

[0063] Long-term trend analysis module, used to perform time series modeling and multi-scale rate of change analysis on preprocessed data;

[0064] Adaptive risk model module, used to set and dynamically adjust risk thresholds based on long-term trend analysis results;

[0065] A space-time data fusion module to fuse data from different sources and generate a geological risk index;

[0066] Abnormal event feedback and adaptive learning module, used to adjust risk threshold strategy based on monitoring result feedback;

[0067] The intelligent early warning and emergency response module is used to generate early warning signals and trigger emergency responses based on geological risk indexes.

[0068] Please see attached Figure 2 , the data acquisition module includes:

[0069] GNSS monitoring equipment, used to collect three-dimensional displacement data;

[0070] InSAR data acquisition equipment, used to collect surface subsidence data;

[0071] Inclinometers, used to collect tilt angle data in mining areas;

[0072] Seismic sensors are used to collect data on microseismic activity.

[0073] Specifically, the data acquisition module of the present invention is used to collect multi-source geological data from multiple monitoring points in the mining area to comprehensively cover the geological change information of the mining area, especially geological parameters such as three-dimensional displacement, surface settlement, tilt angle and microseismic activity. This module is composed of the following main equipment: GNSS monitoring equipment, InSAR data acquisition equipment, inclinometer and seismic sensor. Each device has different functions in the data acquisition module. The specific working principle, technical details and implementation methods are as follows:

[0074] The Global Navigation Satellite System (GNSS) monitoring equipment is based on satellite positioning technology and uses the distance difference between satellite signals and ground receivers to calculate the three-dimensional displacement of the monitoring point. The basic principle of GNSS technology is to receive multiple satellite signals, combine the position information of satellites and receivers, and use triangulation to determine the three-dimensional coordinates of ground monitoring points.

[0075] GNSS equipment mainly consists of a GNSS receiver, an antenna and a data transmission module.

[0076] GNSS receivers collect signals from multiple GNSS satellites (such as GPS, GLONASS, Galileo, and BeiDou) and use the satellite orbit information to calculate the three-dimensional position of the receiver.

[0077] The collected data include real-time three-dimensional displacement (easting, northing and elevation change) with millimeter-level accuracy, which is suitable for detecting surface deformation within the mining area.

[0078] GNSS receivers use differential GPS (DGPS) or real-time kinematic (RTK) technology to improve positioning accuracy, which is particularly critical for high-precision surface subsidence and displacement monitoring.

[0079] Implementation method:

[0080] GNSS equipment is installed at key monitoring points in the mining area. The distribution of monitoring points is determined according to the geological structure and mining activities to cover areas where deformation may occur.

[0081] The device collects data at regular intervals (such as measuring once every hour) and transmits the data to the central server for use by subsequent processing modules.

[0082] The location data of each monitoring point is compared with previous historical data to analyze the three-dimensional displacement caused by mining activities and identify abnormal situations such as slow settlement or sudden displacement.

[0083] Synthetic Aperture Radar Interferometry (InSAR) is a surface deformation monitoring technology based on radar satellite images. InSAR compares radar images of the same area acquired at different times, analyzes the phase difference of radar waves, and calculates the subsidence or uplift of the surface.

[0084] InSAR equipment includes synthetic aperture radar (SAR) satellites and ground receiving systems. SAR satellites periodically acquire radar images of the mining area and obtain phase information of the same location at different times.

[0085] InSAR processes two or more radar images taken at different times, solves the phase difference, generates a differential interferogram, and then calculates the vertical subsidence or uplift of the surface.

[0086] In order to improve accuracy, InSAR uses time-series InSAR (PS-InSAR) technology, which can eliminate interference from factors such as the atmosphere and surface reflection, and achieve surface subsidence monitoring with millimeter-level accuracy.

[0087] Implementation method:

[0088] The system deploys ground receiving devices to regularly receive and store SAR satellite image data and compare it with historical data.

[0089] Phase difference analysis is used to monitor the surface deformation in the mining area, and the subsidence data is periodically updated through InSAR images to obtain the trend of surface subsidence changes.

[0090] InSAR equipment is suitable for large-scale surface deformation monitoring and can complement the point monitoring of GNSS equipment to achieve overall subsidence and uplift monitoring of the mining area.

[0091] Principle: Inclinometers are used to measure changes in the tilt angle of the ground or rock mass, and are mainly based on gravity sensors or acceleration sensors. Their working principle is to determine the change in the tilt angle by detecting changes in the direction of gravity. The inclinometer is equipped with a precise acceleration sensor. When the ground or rock mass tilts, the sensor detects the change in the direction of gravity relative to the device, thereby calculating the tilt angle.

[0092] Inclinometers typically include a dual-axis or tri-axis accelerometer to measure the tilt angle in two or three directions.

[0093] The collected data includes the change in the tilt angle of the monitoring point relative to the horizontal plane (eg, the tilt of the X-axis and the Y-axis).

[0094] The data collection frequency of the inclinometer can be set according to demand, usually set to a higher frequency (such as once every 10 minutes) so that small changes in the surface or rock mass can be detected in time.

[0095] Implementation method:

[0096] The inclinometers are installed at key locations of the rock mass or surface in the mining area, such as slopes and near surface cracks, to ensure coverage of areas where slip may occur.

[0097] The tilt data is transmitted to the monitoring system in real time. The system compares and analyzes the data to identify the changing trend of surface tilt and determine whether there is a risk of landslide or geological instability.

[0098] When the system detects that the inclination exceeds the preset threshold, it triggers an early warning signal, prompting management personnel to conduct on-site inspections and emergency measures.

[0099] Seismic sensors are used to monitor microseismic activity by detecting the vibration of seismic waves through accelerometers or velocimeters. The sensors mainly sense the vibration frequency and amplitude of seismic waves (P waves and S waves) to analyze the stress state and potential rupture risk inside geological structures.

[0100] Seismic sensors include highly sensitive accelerometers that detect microseismic activity in mining areas. Seismic sensors can accurately measure the vibration signals of seismic waves and distinguish between different types of waves (P waves and S waves).

[0101] The collected data include parameters such as frequency, amplitude, and duration of seismic waves, which can reflect the impact of mining activities on geological structures and help identify possible rupture points or fault slips.

[0102] Seismic sensors can monitor micro-earthquakes in real time. They are highly sensitive and can detect tiny vibrations of extremely low magnitudes. They are suitable for monitoring changes in geological stress caused by mining activities.

[0103] Implementation method:

[0104] Multiple seismic sensors are deployed in the mining area to form a microseismic monitoring network to monitor stress changes caused by mining activities.

[0105] The data is transmitted to the monitoring system in real time, which identifies abnormal microseismic activity by analyzing the vibration frequency and amplitude of seismic waves to determine whether there is a potential risk of geological damage.

[0106] When microseismic activity is detected exceeding the preset threshold, the system records the earthquake event, updates the stress state, and promptly reports possible damage risks to the management system.

[0107] In the data acquisition module, GNSS monitoring equipment, InSAR data acquisition equipment, inclinometers and seismic sensors collect different types of geological data. These data are transmitted to the central server of the monitoring system in real time via wireless or wired networks. The system inputs these data into the data preprocessing module for synchronization, alignment, cleaning and standardization to ensure data quality and consistency, and provide high-quality data support for subsequent trend analysis, risk assessment and emergency response.

[0108] Please see attached Figure 3 , the data preprocessing module includes:

[0109] A time synchronization unit, used to align different types of data to a unified time step;

[0110] The spatial alignment unit is used to unify the data of different monitoring points into the same coordinate system;

[0111] Data cleaning unit, used to clean data using quantile regression method and remove outliers;

[0112] Data normalization unit, used to normalize data to the range [-1,1].

[0113] Specifically, different types of data acquisition devices have different data acquisition frequencies and timestamp standards, which will cause timing errors when used directly for analysis. The role of the time synchronization unit is to align various types of data to a unified time step (for example, every hour) to ensure the consistency of the data in the time dimension.

[0114] The time synchronization unit receives data streams from different devices and reads the timestamp information of each data.

[0115] For data of different time series, linear interpolation or spline interpolation method is used to interpolate the lower frequency data to a uniform time step. The formula of linear interpolation is:

[0116]

[0117] Among them, X(t) is the interpolated data value, X1 and X2 are data values ​​at adjacent time points, and t1 and t2 are adjacent time points.

[0118] Implementation method:

[0119] The system sets a uniform time step (such as 1 hour or 30 minutes) so that data from different devices can be synchronized to the same time point.

[0120] For data collected less frequently (such as InSAR data), linear interpolation or spline interpolation is used to smooth it to the required time step.

[0121] The aligned data stream is sent to the next processing unit to ensure that all types of data are consistent at the same time point.

[0122] Different monitoring points may use different coordinate systems (for example, some devices use local coordinate systems, while other devices use geocentric coordinate systems), and a unified coordinate system is required for data analysis. The role of the spatial alignment unit is to map the data of all monitoring points to a unified coordinate system (such as the UTM coordinate system) to ensure consistency in spatial dimensions.

[0123] The spatial alignment unit reads the original coordinate data of each monitoring point and identifies the coordinate system type of the data (eg, geographic coordinate system, local coordinate system, etc.).

[0124] Use the coordinate conversion formula to unify the data of different coordinate systems into the specified target coordinate system (such as UTM coordinate system). The conversion formula is:

[0125]

[0126] Among them, (X, Y, Z) is the coordinate in the target coordinate system, (x, y, z) is the original coordinate, R is the rotation matrix, and T is the translation vector.

[0127] Implementation method:

[0128] The system sets a unified target coordinate system (such as the UTM coordinate system) to ensure that data collected by different devices can be processed in the same spatial framework.

[0129] For data using different coordinate systems, conversion formulas are used to transform the coordinates to ensure the spatial consistency of the data.

[0130] The spatially aligned data will be sent to the next processing unit to ensure the consistency of the data in the spatial dimension.

[0131] In the process of data collection, outliers (such as extreme deviation values ​​or collection errors) will inevitably appear, which will affect the reliability of the data. The data cleaning unit uses the quantile regression method to clean the data and remove abnormal data points through quantile analysis.

[0132] The quantile regression method is a robust anomaly detection method that identifies outliers by analyzing the distribution of data at different quantiles. The optimization formula of the quantile regression model is:

[0133]

[0134] Among them, ρ τ (u) = u(τ-I(u<0)), where τ is the quantile used to adjust the threshold of anomaly detection.

[0135] The system can select different quantiles (such as 90%, 95%, etc.), set different cleaning intensities for different types of data, and eliminate extreme data points.

[0136] Implementation method:

[0137] The system first establishes a quantile regression model for each type of data and calculates the quantile distribution of the data.

[0138] According to actual needs, a quantile threshold (such as 90%) is set, and extreme data points exceeding this threshold are removed.

[0139] The cleaned data is sent to the next processing unit to ensure the reliability of the data.

[0140] The data units and dimensions collected by different devices are different, and directly using them for analysis will lead to inconsistency in model processing. The data standardization unit standardizes each data to the range of [-1,1] to eliminate the difference in data dimensions, so that different types of data can be processed and analyzed under the same framework.

[0141] The data standardization unit uses the minimum-maximum normalization method to convert the data to the [-1,1] interval. The formula is as follows:

[0142]

[0143] Among them, X i,j is the original data value, X min,j and X max,j are the minimum and maximum values ​​of the data column, respectively. i,j is the normalized data value.

[0144] The system first calculates the maximum and minimum values ​​of each type of data to determine the standardized benchmark.

[0145] The original data is substituted into the standardization formula for normalization, and the converted data values ​​are restricted to the range of [-1,1].

[0146] The standardized data is fed into the long-term trend analysis module to ensure that the subsequent analysis model can consistently handle different data sources.

[0147] Please see attached Figure 4 , the long-term trend analysis module includes:

[0148] The space-time joint decomposition unit is used to decompose the data into time trend terms, spatial variation terms and residual terms;

[0149] Wavelet transform multi-scale analysis unit is used to capture geological changes at different time scales.

[0150] Specifically, the geological data of the mining area is affected by both time and space, so decomposing the data into time trend terms, spatial variation terms, and residual terms helps to extract different characteristics of time and space respectively. This decomposition method can reveal the long-term geological change trend under the influence of mining activities and the differences in geological structure within the region.

[0151] Time trend term T(t): represents the long-term trend over time, reflecting the geological deformation trend of the mining area, such as the gradually accumulated surface subsidence.

[0152] Spatial variation term S(x,y): represents the difference in spatial dimension, reflecting the geological structure characteristics of different locations in the mining area, such as differences in rock hardness and stratum composition.

[0153] The residual term R(x,y,t) represents short-term fluctuations and unexplained random components, which is used to capture unpredictable geological events or noise.

[0154] The mathematical model of joint decomposition is:

[0155] X(x,y,t)=T(t)+S(x,y)+R(x,y,t)

[0156] Among them, X(x,y,t) is the observed data value of the mining area at location (x,y) and time t, T(t) is the time trend term, S(x,y) is the spatial variation term, and R(x,y,t) is the residual term.

[0157] Time trend extraction: First, the collected geological data are grouped by time, the average or median of each time period is calculated, the time series is constructed and the long-term trend is fitted.

[0158] Spatial variation extraction: After eliminating the time trend, the spatial data distribution of the mining area is analyzed, and the geological difference characteristics of different locations are extracted through methods such as geographically weighted regression (GWR) or kriging interpolation.

[0159] Residual term calculation: remove the time trend term and spatial variation term from the original data to obtain the residual part. The residual is used to further analyze the impact of emergencies or noise on geological data.

[0160] Result transmission: The decomposed time trend term, spatial variation term and residual term are transferred to the subsequent wavelet transform multi-scale analysis unit as input data.

[0161] Wavelet transform is a mathematical tool for time-frequency analysis that can decompose data into different time scales to capture short-term and long-term geological change characteristics. Through wavelet transform, multi-scale analysis units can identify multi-level characteristics of short-term sudden events (such as landslides and subsidence) and long-term slow changes (such as gradual accumulation of subsidence) in geological data.

[0162] Wavelet transform decomposes time series into components at different scales (frequency bands) and can simultaneously analyze the time and frequency characteristics of the signal.

[0163] Discrete wavelet transform (DWT) is used to decompose the short-term, medium-term and long-term components in the time trend term, enabling the system to identify the change patterns at each time scale separately.

[0164] The wavelet decomposition formula is as follows:

[0165]

[0166] Among them, W ψ (a, b) are wavelet coefficients, X(t) is the input time series data, ψ a,b (t) is the scale

[0167] Wavelet function at position a and b. Wavelet coefficient W ψ (a, b) show the signal strength at different scales and locations.

[0168] Wavelet selection: Select an appropriate wavelet basis function (such as Daubechies wavelet or Haar wavelet) for data decomposition. The choice of wavelet basis is based on the characteristics of the data to ensure that both short-term and long-term variation characteristics can be captured.

[0169] Multi-scale decomposition: Input the time trend item data into the wavelet transform, and perform multi-scale decomposition on the data through DWT to obtain wavelet coefficients on different time scales. The low-frequency component is used to capture long-term trends, while the high-frequency component is used to detect short-term sudden changes.

[0170] Component analysis: Analyze the different scale components after wavelet decomposition. The low-frequency component (long-term trend) is used to identify the slowly accumulated settlement trend; the high-frequency component (short-term fluctuation) is used to identify rapid changes in the short term, such as surface landslides, sudden settlement, etc.

[0171] Result output: The multi-scale analysis results will be passed to the adaptive risk model module to dynamically adjust the risk threshold and provide accurate multi-scale data support for early warning and emergency response.

[0172] Please see attached Figure 5 , the adaptive risk model module includes:

[0173] An adaptive threshold setting unit, used to set an initial threshold based on the long-term trend analysis result;

[0174] The event-driven dynamic threshold update unit is used to dynamically adjust the risk threshold based on event feedback information when a geological anomaly event is detected.

[0175] The dynamic adjustment in the event-driven dynamic threshold update unit is based on the following update formula:

[0176] T new =T prev +λ·ΔX

[0177] Among them, T new is the updated threshold, T prev is the previous threshold, λ is the adjustment coefficient, and ΔX is the change amplitude.

[0178] Specifically, the adaptive risk model module aims to use the geological data collected and analyzed to dynamically adjust the system's risk threshold according to changes in geological risks, so as to achieve real-time response to risks in the mining area. Through the setting and dynamic adjustment of adaptive risk thresholds, the system can make corresponding adjustments according to changes in geological risk conditions, reduce false alarms and missed alarms, and improve the accuracy of early warnings.

[0179] The role of the adaptive threshold setting unit is to set the initial threshold of geological risk based on the geological data trend results extracted by the long-term trend analysis module. The initial threshold reflects the risk level under normal geological conditions and serves as the basis for event-driven dynamic adjustment. This threshold can adapt to the characteristics of different mining areas and ensure the accuracy of the initial risk assessment.

[0180] Initial threshold calculation: The adaptive threshold setting unit sets the initial threshold using the weighted average method based on the mean and standard deviation of historical geological data in order to establish a basic reference for normal geological changes in the mining area. The formula for setting the initial threshold is:

[0181] T0=μ hist +k·σ hist

[0182] Among them, T0 is the initial threshold, μ hist is the mean of historical data, σ hist is the standard deviation of historical data, and k is the adjustment factor used to control the sensitivity of the threshold.

[0183] Threshold adaptability: The system regularly checks and updates the mean and standard deviation of historical data to ensure that the thresholds adapt to the long-term dynamics of geological trends. For example, as mining activity increases or geological structures change, the thresholds are updated accordingly so that they reflect the true state of geological stability.

[0184] Implementation method:

[0185] Data input: Obtain statistical characteristics of historical geological data (such as mean and standard deviation) from the long-term trend analysis module.

[0186] Calculate the initial threshold: According to the set adjustment factor kkk, use the historical mean and standard deviation to calculate the initial threshold T0.

[0187] Storing initial thresholds: The calculated initial thresholds are stored in the system as benchmark values ​​for risk assessment and event-driven dynamic adjustments.

[0188] Periodic Updates: Initial thresholds are recalculated and updated periodically or when significant changes are detected in the system to ensure they adapt to long-term geological changes.

[0189] The event-driven dynamic threshold update unit is used to dynamically adjust the threshold according to the feedback information of the event when a geological anomaly is detected, so as to enhance the adaptability of the system. The dynamic adjustment mechanism makes the system more responsive to changes in geological risks by updating the threshold after an abnormal event occurs.

[0190] Dynamic adjustment formula: Event-driven dynamic threshold update is based on the empirical Bayes formula, and the new threshold T is calculated according to the feedback information of the event.new The specific update formula is:

[0191] T new =T prev +λ·ΔX

[0192] Where: T new is the updated threshold; T prev is the previous threshold; λ is the adjustment coefficient, which controls the amplitude of threshold adjustment; ΔX is the amplitude of change of geological abnormal events (such as the increment of three-dimensional displacement or the change of settlement), which indicates the magnitude of the impact of abnormal events on geological stability.

[0193] Adaptive adjustment of the adjustment coefficient: The system can dynamically adjust the adjustment coefficient λ based on historical data and current risk conditions, increasing the sensitivity of the response in high-risk situations and reducing the sensitivity in low-risk situations to avoid over-adjustment.

[0194] Implementation method:

[0195] Event detection: When the system detects a geological anomaly (such as displacement exceeding the warning threshold or increased microseismic activity), an event-driven dynamic threshold update process is triggered.

[0196] Calculate the change amplitude: Based on the current monitoring data, calculate the change amplitude ΔX of the abnormal event, for example, calculate the increment between the current displacement and the previous displacement.

[0197] Dynamically adjust the threshold: Use the dynamic adjustment formula, substitute the change range ΔX into the formula, and calculate the new risk threshold T new .

[0198] Store the updated result: the updated threshold T new Stored as a new baseline threshold for subsequent risk assessment.

[0199] Adaptive adjustment coefficient: After a period of time, the effect of the adjustment coefficient λ is evaluated based on historical event data, and adjustments are made when necessary to improve the accuracy of risk assessment.

[0200] Workflow Integration:

[0201] Initial threshold setting: When the system starts, the adaptive threshold setting unit receives the historical data provided by the long-term trend analysis module, calculates and stores the initial threshold T0T_0T0 as a reference benchmark under normal risk conditions.

[0202] Event-driven threshold adjustment: During the monitoring process, the event-driven dynamic threshold update unit monitors the occurrence of abnormal events in real time. When an abnormal event is detected, the system dynamically adjusts the risk threshold T based on the feedback information. new .

[0203] Cyclic update: The thresholds are continuously adjusted as new geological events occur, allowing the system to adapt to the dynamic changes in geological risks and avoid false positives or negative positives when the environment changes.

[0204] The space-time data fusion module uses Gaussian process regression method to fuse different data sources to generate the geological risk index I of the mining area. risk , the Gaussian process regression method is based on the following formula:

[0205] f(x)~GP(m(x),k(x,x′))

[0206] Among them, m(x) is the mean function, which is used to represent the average value of the overall geological stability, and k(x,x ′ ) is the covariance function, which is used to reflect the spatial-temporal similarity.

[0207] Specifically, the main function of the space-time data fusion module is to fuse geological monitoring data from different data sources (such as GNSS displacement data, InSAR settlement data, tilt angle data and seismic sensing data) to generate a comprehensive geological risk index to quantify the geological stability of the mining area. To this end, the module uses the Gaussian process regression (GPR) method to capture the correlation between data sources through the covariance function of space and time, thereby achieving effective fusion of different data.

[0208] Gaussian process regression is a non-parametric Bayesian regression method. Its core idea is to regard the input data as a set of random variables in a high-dimensional space. By defining the mean function and covariance function, the correlation between known data points is used to predict unobserved points. The advantage of GPR in space-time data fusion is that it can describe the smooth changes of data in space and time dimensions through the covariance function.

[0209] Gaussian process regression model formula:

[0210] Mean function m(x): describes the average trend of data at different locations and times. The mean function is defined as:

[0211] m(x)=E[f(x)]

[0212] Among them, f(x) is the measured value of the geological data of the mining area at location and time, and the mean function reflects the average value of the overall geological stability.

[0213] Covariance function k(x,x ′ ): used to describe data points x and x ′ The correlation between them, the covariance function is defined as:

[0214] k(x,x′)=E[(f(x)-m(x))(f(x′)-m(x′))]

[0215] The covariance function not only considers the spatial similarity of the data, but also introduces the dependency of the time dimension, thereby achieving the modeling of the spatial-temporal similarity of different monitoring points in the mining area.

[0216] The Gaussian process regression model can finally be expressed as:

[0217] f(x)~GP(m(x),k(x,x′))

[0218] Where GP represents Gaussian process, m(x) is the mean function, and k(x,x ′ ) is the covariance function.

[0219] Choice of covariance function:

[0220] In Gaussian process regression, the choice of covariance function determines how spatial and temporal correlations are modeled. Commonly used covariance functions include the following:

[0221] Squared exponential covariance function:

[0222]

[0223] Among them, σ 2 is the signal variance, l is the length scale, and controls the decay rate of spatial similarity. The square exponential function is suitable for smoothly changing geological data.

[0224] Mattern covariance function:

[0225]

[0226] Among them, ν controls the smoothness, K ν is the Bessel function. The Mattern covariance function can flexibly adjust the smoothness and is suitable for geological data with mutations.

[0227] Space-time composite covariance function: By combining the separate covariance functions of space and time, a composite covariance function can be obtained, for example:

[0228] k((x,t),(x′,t′))=k space (x,x′)·k time (t,t′)

[0229] Among them, k space (x,x ′ ) represents the covariance function in space, k time (t,t ′) represents the temporal covariance function. This composite function is suitable for considering correlations in both spatial and temporal dimensions.

[0230] Implementation method:

[0231] Data input: The space-time data fusion module receives geological monitoring data from different data sources, including displacement, settlement, tilt angle, microseismic activity, etc. These data have different spatial locations and time stamps.

[0232] Setting of mean function: First, conduct exploratory analysis on the data to determine the overall geological stability trend and establish the mean function m(x). The mean function can be set to zero mean or based on the average trend of historical data.

[0233] Selection and optimization of covariance function: Select an appropriate covariance function (such as squared exponential or Mattern covariance function) based on the data characteristics, and determine the hyperparameters of the covariance function (such as length scale l and signal variance σ) by maximum likelihood estimation or Bayesian optimization methods.

[0234] Constructing a Gaussian process model: Applying the mean function and covariance function to the data, constructing a Gaussian process model f(x)~GP(m(x),k(x,x ′ )). The Gaussian process model can be used to make predictions for unobserved points and generate geological risk estimates for each location and time point in the mining area.

[0235] Generate geological risk index: Use the results predicted by the Gaussian process regression model to generate a comprehensive geological risk index I for the entire mining area. risk , which quantifies the geological stability status of the mining area. The geological risk index can be expressed as the predicted mean value and its uncertainty for each monitoring point, which is used for subsequent early warning and risk assessment.

[0236] Result output: The generated geological risk index is transmitted to the intelligent early warning and emergency response module to support the system's risk warning and emergency decision-making.

[0237] Please see attached Figure 6 , the abnormal event feedback and adaptive learning module includes:

[0238] An abnormal event detection unit, used to detect abnormal events based on local abnormal factors;

[0239] The adaptive feedback learning unit is used to use the Q-learning reinforcement learning method to adjust the risk threshold strategy according to the feedback of the detected abnormal events.

[0240] The wavelet transform multi-scale analysis unit performs wavelet decomposition based on the following formula:

[0241]

[0242] Among them, W ψ (a,b) are wavelet coefficients, a is the scale factor, and b is the translation factor.

[0243] Specifically, the abnormal event detection unit detects abnormal geological events in the mining area through the local anomaly factor algorithm. The LOF algorithm is a density-based anomaly detection method that determines whether a data point is an abnormal point by comparing the local density of the data point with its neighboring points. LOF can effectively identify outliers in the data and is particularly suitable for detecting sudden anomalies in geological monitoring data.

[0244] LOF calculation formula:

[0245]

[0246] Where: A represents the target point to be detected; B i represents the i-th neighboring point of A; k is the number of neighboring points; lrd(A) is the local reachability density of the target point, which is defined as the inverse of the average distance between A and its neighboring points. If the LOF value of the target point A is significantly greater than 1, its density is lower than the density of the surrounding points, indicating that A is an outlier.

[0247] Implementation method:

[0248] Data input: The abnormal event detection unit receives geological monitoring data from the space-time data fusion module.

[0249] Neighboring point selection: For each monitoring point A, select k nearest neighboring points B i , to calculate the local density.

[0250] Calculate LOF value: Calculate the LOF value of each monitoring point according to the LOF formula.

[0251] Abnormal judgment: The LOF value is compared with the set abnormal threshold. If the LOF value exceeds the threshold, the monitoring point is marked as an abnormal point and its information is passed to the adaptive feedback learning unit.

[0252] The role of the adaptive feedback learning unit is to dynamically adjust the risk threshold strategy based on the feedback of detected abnormal events. The unit uses the Q-learning reinforcement learning algorithm to continuously optimize the threshold setting of the system by learning the abnormal patterns of geological data. Q-learning is a model-free reinforcement learning algorithm that learns the best strategy through trial and error. In this system, Q-learning is used to adjust the threshold when the geological risk changes to reduce false positives and false negatives.

[0253] Q-learning algorithm formula:

[0254] The core of Q-learning is the Q value update formula:

[0255]

[0256] Where: Q(s,a) is the Q value of executing action a in the current state s; α is the learning rate, which controls the step size of Q value update; r is the immediate reward after executing the action, which is used to reflect the effect after adjusting the threshold; γ is the discount factor, which is used to control the impact of future rewards; s ′ is the next state after executing action a, a ′ is the best action in the next state.

[0257] Implementation method:

[0258] Definition of states and actions:

[0259] Status s: defined according to the current geological risk status (e.g. the number of anomalies detected by LOF).

[0260] Action a: includes operations such as "increase threshold", "decrease threshold", and "maintain current threshold" to adjust the risk response of the system.

[0261] Q value initialization: Initialize the Q values ​​of all state-action pairs, set the initial learning rate α and discount factor γ.

[0262] Strategy selection: After each abnormal event detection, select an action aaa according to the current state sss, such as "increase threshold" or "decrease threshold" to adjust the risk response.

[0263] Q value update: After executing the selected action, calculate the immediate reward r according to the feedback result and update the Q value, and record the adjustment effect to optimize the threshold strategy.

[0264] Iterative learning: Through continuous iteration, each abnormal event detection and feedback loop can gradually optimize the Ql earning strategy, allowing the system to adaptively adjust thresholds when geological risks change.

[0265] Wavelet transform multi-scale analysis formula description:

[0266] The wavelet transform multiscale analysis unit is used to capture the multiscale characteristics of geological data, using the following wavelet decomposition formula:

[0267]

[0268] in:

[0269] W ψ(a, b) are wavelet coefficients; a is the scale factor, which controls the resolution of wavelet decomposition; b is the translation factor, which controls the position of the wavelet; X(t) is the input time series data; ψ a,b (t) is the wavelet basis function, which is used to decompose signals at different scales and locations.

[0270] The results of wavelet transform are signal characteristics at different scales, which are helpful to identify short-term and long-term geological change trends. The results of wavelet transform multi-scale analysis are used to further detect abnormal events and provide data support for adaptive feedback learning.

[0271] Please see attached Figure 7 , the intelligent early warning and emergency response modules include:

[0272] a risk assessment unit for receiving the geological risk index and comparing it with a dynamic threshold to determine the geological risk level;

[0273] A multi-level warning generation unit is used to generate corresponding warning signals according to the geological risk level;

[0274] An emergency decision-making unit, used to generate an emergency response plan corresponding to the warning signal;

[0275] Dynamic emergency response optimization unit, used to adjust emergency response plans based on actual geological change information;

[0276] Emergency plan simulation and drill unit, used to simulate and drill emergency plans;

[0277] The log recording and report generation unit is used to record information during the early warning and emergency response process and generate event reports.

[0278] Specifically, the risk assessment unit receives the geological risk index generated by the space-time data fusion module, and compares the index with the system's dynamic threshold to determine the current geological risk level. The unit can classify risks into different levels, such as low risk, medium risk, and high risk, according to the size of the risk index, thereby providing a basis for subsequent warning generation.

[0279] Implementation method:

[0280] Receiving data: The risk assessment unit receives the real-time updated geological risk index from the space-time data fusion module.

[0281] Threshold comparison: Compare the geological risk index with the dynamic threshold to determine the current risk level. For example, a risk index below the threshold is considered low risk, while one above the threshold is considered medium or high risk.

[0282] Output risk level: Output the corresponding geological risk level based on the comparison results and pass it to the multi-level warning generation unit to activate the corresponding warning signal.

[0283] The multi-level warning generation unit generates different levels of warning signals, such as general warning, warning and emergency warning, according to the output risk level of the risk assessment unit. The purpose of the multi-level warning is to allow managers and systems to take appropriate response measures according to the risk level.

[0284] Implementation method:

[0285] Receive risk level: Receive the geological risk level output by the risk assessment unit.

[0286] Generate early warning signals: Generate corresponding early warning signals according to the risk level. For example:

[0287] Low risk: General warning, prompting system monitoring.

[0288] Medium risk: Warning, reminding managers to pay close attention.

[0289] High risk: Emergency warning, prompting urgent response measures.

[0290] Warning signal notification: Warning signals are sent to relevant managers and systems via SMS, email or on-site alarm system for timely response.

[0291] The emergency decision-making unit formulates a corresponding emergency response plan based on the generated warning signal. Different warning levels correspond to different emergency strategies. For example, in high-risk situations, it may be necessary to immediately evacuate personnel and stop work, while in medium-risk situations, it may only be necessary to add monitoring forces.

[0292] Implementation method:

[0293] Receiving warning signals: The emergency decision-making unit receives warning signals from multi-level warning generation units.

[0294] Develop emergency response plans: Develop emergency response plans based on the warning level, such as:

[0295] General warning: Increase monitoring frequency.

[0296] Warning: Send personnel to conduct on-site inspection and prepare emergency measures.

[0297] Emergency warning: Initiate emergency responses such as personnel evacuation, shutdown and equipment protection.

[0298] Implementation plan notification: Notify relevant departments and personnel of the emergency response plan to ensure that emergency measures are implemented.

[0299] The dynamic emergency response optimization unit is used to adjust the original emergency response plan according to the real-time monitored geological change information during the emergency response process. The unit can optimize the emergency strategy during the response implementation process to ensure that the response plan adapts to the real-time situation of geological changes.

[0300] Implementation method:

[0301] Monitor real-time changes: The dynamic emergency response optimization unit receives geological monitoring data in real time to determine the adaptability of the current emergency response plan.

[0302] Optimization and adjustment: If the monitoring data shows that the geological conditions have further deteriorated or improved, the emergency response plan can be adjusted dynamically. For example, if the risk index increases rapidly, the response measures can be strengthened; if the risk index decreases, the response intensity can be appropriately slowed down.

[0303] Adjustment feedback: The adjusted emergency plan will be fed back to the emergency decision-making and execution department to ensure that the response measures are in line with the actual situation of the current geological risks.

[0304] The emergency plan simulation and drill unit is used to simulate and drill the emergency plan in non-emergency situations to verify the effectiveness of the emergency response plan. Through simulation and drill, the system can identify the weak links of the emergency plan and make improvements in advance.

[0305] Implementation method:

[0306] Plan setting: The emergency plan simulation and drill unit sets emergency plans for different scenarios, including low, medium and high risk emergency response measures.

[0307] Simulation execution: In non-emergency situations, simulate different response measures in the plan, including personnel evacuation, equipment shutdown and other operations, to check the feasibility and effectiveness of the plan.

[0308] Record the results of the drill: record the execution effect and timing of each measure during the simulation process to identify weak points or delays in the process.

[0309] Improve the plan: Adjust the emergency plan based on the simulation results to ensure the best execution effect in the real emergency response.

[0310] The log recording and report generation unit is used to automatically record the time, operation content and execution results of each operation link during the early warning and emergency response process, and generate a complete event report. This unit provides detailed data support for subsequent risk assessment and system improvement.

[0311] Implementation method:

[0312] Record operation logs: During the early warning and emergency response process, each operation is automatically recorded, including early warning signal generation, emergency decision-making plan, response plan adjustment, etc.

[0313] Generate report: After the event is over, summarize all operation records and generate an event report, including detailed information such as warning trigger time, response measures, adjustment process, execution results, etc.

[0314] Archiving and Analysis: Archive incident reports for subsequent risk assessment and system improvement analysis to identify successful experiences and room for improvement in emergency response processes.

[0315] Working principle: First, the data acquisition module uses a variety of sensor equipment such as GNSS monitoring equipment, InSAR data acquisition equipment, inclinometers and seismic sensors to collect data such as three-dimensional displacement, surface settlement, tilt angle and microseismic activity from multiple monitoring points in the mining area. These data provide a comprehensive source of geological information for subsequent analysis.

[0316] Next, the collected multi-source data is normalized through the data preprocessing module. The time synchronization unit of this module aligns the data from different sources to a unified time step to ensure temporal consistency; the spatial alignment unit unifies the data of each monitoring point into the same coordinate system to ensure spatial consistency. Subsequently, the data cleaning unit uses the quantile regression method to clean the data, remove outliers, and ensure the reliability of the data. Finally, the data standardization unit standardizes each data to the [-1,1] interval to eliminate differences in data dimensions and make the data suitable for subsequent analysis and modeling.

[0317] The preprocessed data are input into the long-term trend analysis module, where the space-time joint decomposition unit decomposes the data into time trend terms, spatial variation terms and residual terms to extract the time change trend, spatial distribution characteristics and short-term fluctuations of the data respectively. Then, the wavelet transform multi-scale analysis unit further decomposes the time trend terms into multiple scales to capture geological changes at different time scales, enabling the system to identify short-term and long-term geological risk characteristics and adapt to complex geological dynamics.

[0318] After obtaining the results of the long-term trend analysis, the system enters the adaptive risk model module. The adaptive threshold setting unit in this module sets the initial threshold based on the results of the trend analysis and defines the initial judgment criteria for the geological risk in the mining area. When an abnormal geological event is detected, the event-driven dynamic threshold update unit dynamically adjusts the risk threshold by analyzing the feedback information of the event, so that the threshold can adapt to geological changes in real time. This dynamic adjustment uses the empirical Bayesian formula to ensure the flexibility of the threshold and avoid missed reports and false reports.

[0319] In order to further integrate the information from various data sources, the space-time data fusion module uses the Gaussian process regression method to fuse data from different sources into a comprehensive geological risk index. This index combines geological information in space and time, accurately reflects the overall geological stability of the mining area, and provides key indicators for risk assessment and emergency response.

[0320] When the system identifies geological risks through risk index, the abnormal event feedback and adaptive learning module will detect and learn the abnormal events. The abnormal event detection unit of this module uses the local abnormal factor algorithm to identify potential abnormal events. At the same time, the adaptive feedback learning unit adopts the Ql earning reinforcement learning method to dynamically adjust the risk threshold strategy according to the detected abnormal event feedback, and continuously optimize the response accuracy of the system.

[0321] According to the geological risk index and threshold judgment, the system finally enters the intelligent early warning and emergency response module. This module first compares the risk index with the threshold through the risk assessment unit to determine the geological risk level. The multi-level early warning generation unit generates corresponding early warning signals according to the risk level, such as general early warning, warning, and emergency early warning, and sends reminders to managers or systems.

[0322] After the warning is generated, the emergency decision-making unit will formulate an emergency response plan that matches the warning level, including specific measures such as personnel evacuation and equipment shutdown. The dynamic emergency response optimization unit monitors geological changes in real time and adjusts the emergency plan to ensure that the emergency response strategy adapts to the current risks. The system also simulates and exercises the emergency plan in non-emergency situations through the emergency plan simulation and drill unit to ensure the effectiveness of the plan.

[0323] During the entire monitoring and emergency response process, the log recording and report generation unit will automatically record the time, operation and results of each response link, generate a complete event report, and provide data support for subsequent risk analysis and system optimization.

[0324] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A dynamic monitoring system for geological stability in mining areas, characterized in that: include: A data acquisition module for collecting geological monitoring data from multiple data sources within the mining area; Data preprocessing module, used for time synchronization, spatial alignment, data cleaning and standardization of collected data; Long-term trend analysis module, used to perform time series modeling and multi-scale rate of change analysis on preprocessed data; Adaptive risk model module, used to set and dynamically adjust risk thresholds based on long-term trend analysis results; A space-time data fusion module to fuse data from different sources and generate a geological risk index; Abnormal event feedback and adaptive learning module, used to adjust risk threshold strategy based on monitoring result feedback; The intelligent early warning and emergency response module is used to generate early warning signals and trigger emergency responses based on geological risk indexes.

2. The dynamic monitoring system for geological stability of mining areas according to claim 1, characterized in that: The data acquisition module comprises: GNSS monitoring equipment, used to collect three-dimensional displacement data; InSAR data acquisition equipment, used to collect surface subsidence data; Inclinometers, used to collect tilt angle data in mining areas; Seismic sensors are used to collect data on microseismic activity.

3. The dynamic monitoring system for geological stability of mining areas according to claim 1, characterized in that: The data preprocessing module comprises: A time synchronization unit, used to align different types of data to a unified time step; The spatial alignment unit is used to unify the data of different monitoring points into the same coordinate system; The data cleaning unit is used to clean the data and remove outliers using the quantile regression method; Data normalization unit, used to normalize data to the range [-1,1].

4. The dynamic monitoring system for geological stability of mining areas according to claim 1, characterized in that: The long-term trend analysis module includes: The space-time joint decomposition unit is used to decompose the data into time trend terms, spatial variation terms and residual terms; Wavelet transform multi-scale analysis unit is used to capture geological changes at different time scales.

5. The dynamic monitoring system for geological stability of mining areas according to claim 1, characterized in that: The adaptive risk model module includes: An adaptive threshold setting unit, used to set an initial threshold based on the long-term trend analysis result; The event-driven dynamic threshold update unit is used to dynamically adjust the risk threshold based on event feedback information when a geological anomaly event is detected.

6. The dynamic monitoring system for geological stability of mining areas according to claim 5, characterized in that: The dynamic adjustment in the event-driven dynamic threshold update unit is based on the following update formula: T new =T prev +λ·ΔX Among them, T new is the updated threshold, T prev is the previous threshold, λ is the adjustment coefficient, and ΔX is the change amplitude.

7. The dynamic monitoring system for geological stability of mining areas according to claim 1, characterized in that: The space-time data fusion module uses Gaussian process regression method to fuse different data sources to generate the geological risk index I of the mining area. risk , the Gaussian process regression method is based on the following formula: f(x)~GP(m(x),k(x,x′)) Among them, m(x) is the mean function, which is used to represent the average value of the overall geological stability, and k(x,x ′ ) is the covariance function, which is used to reflect the spatial-temporal similarity.

8. The dynamic monitoring system for geological stability of mining areas according to claim 1, characterized in that: The abnormal event feedback and adaptive learning module includes: An abnormal event detection unit, used for detecting abnormal events based on local abnormal factors; The adaptive feedback learning unit is used to use the Q-learning reinforcement learning method to adjust the risk threshold strategy according to the feedback of the detected abnormal events.

9. The dynamic monitoring system for geological stability of mining areas according to claim 4, characterized in that: The wavelet transform multi-scale analysis unit performs wavelet decomposition based on the following formula: Among them, W ψ (a,b) are wavelet coefficients, a is the scale factor, and b is the translation factor.

10. The dynamic monitoring system for geological stability of mining areas according to claim 1, characterized in that: The intelligent early warning and emergency response module includes: a risk assessment unit for receiving the geological risk index and comparing it with a dynamic threshold to determine the geological risk level; A multi-level warning generation unit, used to generate corresponding warning signals according to the geological risk level; An emergency decision-making unit, used to generate an emergency response plan corresponding to the warning signal; A dynamic emergency response optimization unit, used to adjust the emergency response plan according to actual geological change information; Emergency plan simulation and drill unit, used to simulate and drill emergency plans; The log recording and report generation unit is used to record information during the early warning and emergency response process and generate event reports.

Images